Wardhani Utami Dewi, Khoirin Nisa, Eri Setiawan, Rizky Ahmad Yudanegara
A multivariate statistical analysis technique that can be used to analyze a structural relationship between latent variables is structural equation modeling (SEM). Variance-based SEM is a solution to the deficiency of covariance-based SEM, assuming that the sample size does not have to be large, the data must be normally distributed, and the indicators can be reflective. One of the variance-based estimation methods for structural modeling is generalized structured component analysis (GCSA). A key benefit of the GSCA method is that it can achieve satisfactory model fit even when with relatively small samples. This is possible because it uses a parameter estimation method called alternating least squares. In this study, we used SEM-GCSA for analyzing the degree of public health in Indonesia in 2020 with a sample size of 34 provinces. The data consists of 21 indicator variables constructing 6 latent variables. The analysis results showed that the variation in the degree of health variables could be explained by 78% of behavioral and environmental variables and the remaining 22% explained by other factors outside the data. The results of the overall model evaluation based on the goodness of fit index values indicate that the model has a good overall fit. © 2024 Author(s).
Department of Mathematics, Faculty of Mathematics and Natural Sciences, Universitas Lampung, Jl. Prof. Dr. Sumantri Brojonegoro No. 1, Lampung, Bandar Lampung, 35145, Indonesia; Department of Infrastructure and Regional Technology, Institut Teknologi Sumatera, Jl. Terusan Ryacudu, Desa Way Hui, Jatiagung, Lampung Selatan, Lampung, 35365, Indonesia